Re: [Accessibility] Call to Arms

[Eric S. Johansson]

 On 7/21/2010 11:16 AM, Chris Hofstader wrote:
> A lot of us, including me, complain about performance and functionality in 
> various applications when running orca. We've got to stop whining and get 
> hacking. It would be good to have people take the lead on Firefox, OpenOffice 
> and other "high value" programs. It would also be useful for people to take 
> charge of entirely new AT (like speech recognition) projects to help us move 
> forward in areas that are not already covered.

I will raise my hand on the speech recognition project but, it comes with some 
caveats.

I believe strongly that the tools first approach you and others have spoken of 
misses the needs of the upper extremity disabled.   their primary need is 
income. You can't have freedom of choice if you can't make money. For example, 
today, if I want to make money, I must use NaturallySpeaking. There is no choice 
and the speech recognition projects available today or the near future are not 
sufficient to replace NaturallySpeaking (I.e. they couldn't write this e-mail 
and they take way too much time to set up).

I would propose organizing the project to first satisfy the economic needs of 
the disabled community, so they can make money, they can be independent and as a 
result, be able to make choices about software freedom.

In the speech recognition domain, that means creating better user interfaces, 
tools to assist in specific tasks, multimachine dictation targets, and also 
changing existing tools (languages, build tools etc.) to work with speech 
recognition instead of trying to make speech recognition work with it. (I.e. 
rebuilding make syntax to something you can speak instead something that is 
convenient for machine)

Once we have built tools, from the outside in, that experience will guide the 
development of OSS speech recognition engines and the eventual replacement of 
systems like NaturallySpeaking.

This philosophy is analogous to the strategy used by the free software 
foundation in its early days.  The GNU toolchain was built to run on top of many 
platforms and this enable people to create the OS and the kernel which gave 
people the power to make choices about free software or not.

Taken same philosophy to speech recognition, I would recommend the following 
path:

Cultivate resources to put NaturallySpeaking under wine. Itit's very close to 
ready and the final push to make it real isn't happening. My reason for this 
step is that it removes all nonfree software except the speech recognition 
engine. This gives us a place to work from, to experiment with various 
techniques for cross machine speech recognition work.

Develop advanced dictation box which is the simplest model for reliable data 
injection into applications.  This is the first layer that will enable upper 
extremity disabled users to make money. This is where you can start writing 
documentation, e-mails etc.

Develop advanced UI techniques such as "interrupting cow" grammars and advanced 
dictation box input and output filters.  this is where system administrators can 
start making money because the grammar would enable a lot of administrator tasks.

Develop programming by voice editing and code creation. We'll probably need a 
lot of help from Emacs folks because we will need to dig down deep into the IDE 
layers.   This is where programmers can start making money because, obviously, 
they can write code. This is also the layer which will enable disabled users to 
contribute to the project. I can't write code using an unenhanced speech 
recognition product.

In theory you can do these in parallel but my best thoughts (so far) lead me to 
the conclusion that you want to use the dictation box with enhancements to 
generate code from the natural output of a large vocabulary continuous speech 
recognition environment.

I consider this amalgam of free and nonfree software acceptable. Instead of 
serving the philosophical and ideological goals all the free software 
foundation, it serves a higher purpose. It enables disabled people to be 
self-sufficient which in turn improves their self-esteem and quality of life. 
When I was disabled in 1994, I was forced to spend 3 1/2 months sitting at my 
desk waiting for them to fire me and receive daily abuse by the managers of the 
company. It took me a long time to get over that and sometimes I think I still 
haven't. Anything we can do to save people that kind of abuse or give them the 
freedom to leave that abuse and go to another company where they are appreciated 
is a very important goal to me.

If you do decide to have me at the core of this project, I will warn you that my 
time is limited as I am currently looking for work. Yes, being disabled even 
with speech recognition gets in the way. I would love to do this kind of thing 
for a living the only folks have found so far would require me to give this IP 
to them. I can't.  it is too important to be held by any one company. This has 
to be a public resource available to anyone.


--- eric